VLIW code generation for LLVM backend


It seems the only one VLIW target Hexagon in LLVM 3.2 devel uses a
straightforward way to emit its VLIW-style asm codes.
It uses a list scheduler to schedule on DAG and a simple packetizer to
wrap the emitted asm instructions.
Both scheduling and packetizing work on basic blocks.

so, is there any plan to implement better optimization methods such as
trace scheduling, software pipelining, ...
or is it already going on?

Best regards.


  There is work currently underway to add SW pipelining and some sort of
global scheduling to Hexagon, but if there is some interest to it from other
targets, it would be helpful to know. What is your involvement with this?

Sergei Larin


Thank you for telling me about this.

Our lab is planning to design a VLIW DSP and has to make a choice
between GCC and LLVM, for which I take responsibility.
As we all know that GCC's codes possess a long history and has a
somewhat bad learning curve, I suggest choosing LLVM.

It seems now the only drawback is its poor support for VLIW
architecture. And so if we can count on LLVM's support for VLIW in a
near future, we will choose LLVM as the compiler infrastructure
without the least hesitation.

Can I do any help for this underway work?



  This might not be such a tough choice on engineering side - one of the
LLVM differentiators is the ground-up, early introduced support for VLIW
specific features...

As for the help offer - thanks... I'll definitely keep that in mind. The
best help right now would be to keep on trying new VLIW related features as
they got merged to LLVM tree and provide feedback... and we do plan to
contribute a lot in the next several weeks/months. Also please participate
in ongoing design discussions - we do want broad audience and quality



Actually, LLVM lacked support for VLIW until fairly recently, and it has
relatively few VLIW-specific features.


But we are really striving to change it fast :slight_smile:


  My comparative point was that LLVM has a greater potential to accommodate
VLIW going forward than GCC does at this point.
Would you agree to that?


Yes, I can see that.
Thank you, Sergei, for the great work.